Advanced Molecular Simulation Methods in the Physical Sciences

Date :From 2013-06-10 To 2013-07-05
Advisory committee :
Local coordinators :Haiping Fang, Zhongcan OuYang, Qiang Shi, Long Wang, Yanting Wang
International coordinators :Christoph Dellago, Yiqin Gao,Yanting Wang, Ruhong Zhou,Xin Zhou
Because molecular simulation methods can provide atomic-scale spatial resolution and femto-second time resolution, they have been recognized as the third pillar of physical sciences besides theory and experiment. However, with the fast progress of scientific research, many physical problems are far beyond the reach of conventional molecular simulation methods. New algorithms based on novel statistical mechanics principles are therefore required to greatly extend the spatial and temporal scales accessible to molecular simulation. This program will mainly focus on the following five topics:

a) Trajectory space sampling. Conventional molecular simulation methods perform importance sampling in configuration space (or phase space), in which a point represents a particular state of the system. For some complex problems, it is advantageous to sample trajectory space, in which each point represents a complete dynamical trajectory. Recently, several new approaches, including the transition path sampling method, have been introduced to improve the sampling efficiency of path-based simulation methods and to extract information on the transition mechanism from the harvested trajectories. These new developments as well as potential applications of trajectory space sampling methods will be discussed in this program.
b) Weighted ensemble sampling. At low temperatures, when a system is trapped in a local minimum of the potential energy surface, it takes a very long time to cross high (free) energy barriers, and therefore conventional molecular simulation methods have difficulties in sampling configuration space ergodically. Advanced simulation methods, such as several versions of replica exchange methods, can be very helpful in alleviating this problem and sample all important configurations in the time scale accessible to the simulation.
c) Coarse-graining methods. All-atom molecular simulation methods treat each atom as an individual particle, which generally limits the spatial size of the simulated system to a length scale of less than tens of nanometers. However, the characteristic length scales for many interesting systems, especially biomolecular systems, often exceeds this limit. Fortunately, most problems related to such systems usually do not require a full resolution at the atomistic scale. Thus coarse-graining (CG) methods, in which groups of several atoms are represented by one particle, are very helpful for those applications. Developing efficient CG methods is very important not only for extending the range of applicability of molecular simulation techniques, but also for developing novel theoretical toolsto study physical phenomena on mesoscopic scales from micrometers to millimeters in size. Current CG methods all have merits and shortcomings, and a general theoretical CG framework does not yet exist. This program will provide a platform for scientists working on the CG methods to communicate and to inspire each other.
d) Free energy calculations. The free energy landscape, rather than the energy landscape, determines the physical properties of a system in equilibrium. Unfortunately, free energies are difficult to compute in molecular simulation results because their calculation often requires the sampling of unlikely configurations. Besides several relatively well-developed techniques, such as the particle insertion method, multiple histograms, acceptance ratio method, and umbrella sampling, the free energy difference between two equilibrium states can also be calculated by a fast non-equilibrium molecular simulation process based on the fluctuation theorems. The discussion of this topic will be conducted to be more focused on the free energy calculations of biomolecular systems.
e) Molecular simulation of self-assembly. Self-assembly has been at the center of numerous studies in recent years, because it is the essential principle in the synthesis of many materials. Some molecular simulation studies have been done for aggregation and nucleation, but very few molecular simulations have been conducted to study self-assembly processes, partly because the interactions between building blocks are very weak and the spatial scale to be simulated is very large. Novel statistical mechanics theories need to be developed specifically for self-assembly and related molecular simulation methods may be developed based on them.
In summary, with five focused topics, this program will provide a platform for scientists from all over the world working on the development of advanced molecular simulation techniques to communicate and inspire each other with new ideas.


Preliminary Schedule:
1) The third week(Jun 24-28) will be scheduled as the dense week with many talks all day long.
2) The fouth week(Jul 1- 5) will be overlapped with the program organized by Liu, Fei and Zhou, Haijun, and it will also be the Satellite Conference of the International Statistical Physics Conference. Haijun is on charge of the Satellite Conference. People from our program are encouraged to attend and give talks on that conference.
3) Participants are encouraged to write review papers about the program topics and submit to the special issue of the journal "Communications in Theoretical Physics".